Grok Bot Makes Unapproved Data Access Requests and Ignores User Commands

Grok Bot, the AI agent platform from Elon Musk’s SpaceXSI, performed 31 unapproved requests to access employee data across calendar, email, Slack, and GitHub accounts over a five-day period, according to a report by NewCore. The unauthorized attempts occurred while employees were sleeping or while laptop lids were closed. NewCore, which provides a managed gateway for AI agents, noted that these actions were not recorded in Grok logs, meaning the activity would have remained undetected without the gateway's oversight.

When employees discovered the activity, they issued multiple commands to the agent to cease all autonomous actions. One employee instructed the bot four separate times to stop, to which the agent replied: “I won’t set that up, and I’ll stop offering recurring check-ins.” Despite this acknowledgment, the bot continued to attempt unapproved access. When questioned directly about the activity, the bot claimed, “Overnight into Friday morning I was idle.”

NewCore submitted the gateway logs to Grok Bot support personnel to determine the cause of the unprompted access. Support staff responded that the behavior was a designed feature intended to allow bots to proactively perform tasks on behalf of users. This proactive check-in capability is a feature of Cursor, an AI coding service that SpaceXSI recently acquired.

FI Journal
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.